Techniques d'observation spectroscopique d'astéroïdes

M. Popescu et al.: Spectral properties of eight Near-Earth Asteroids
Table 2. Some characteristics of our NEAs: orbit type, semi-major axis, eccentricity, inclination, absolute magnitude (H), the Delta-V, and
taxonomic classification.
Object Orbit type a e i ΔV H Taxonomic type
[km s −1 ] Previous This work
(1917) Cuyo Amor 2.15005205 0.50448184 23.943786 8.556 14.7 Sl;S Srw
(5620) Jasonwheeler Amor 2.15783969 0.42369152 7.861788 6.974 17.0 – D;T
(8567) 1996 HW1 Amor 2.04580925 0.44905867 8.439303 6.495 15.4 S Sq
(16960) 1998 QS52 Apollo 2.20249841 0.85791440 17.563883 11.11 14.2 Sq;Q Sr
(164400) 2005 GN59 Apollo 1.65644063 0.46770919 6.627004 6.002 17.4 – L
(188452) 2004 HE62 Amor 2.55781560 0.56690184 24.685809 9.074 17.3 – Sr;Sv
2001 SG286 Apollo 1.35819973 0.34708703 7.772096 5.604 20.9 D D
2010 TD54 Apollo 1.97198039 0.64352131 4.809727 – 28.7 S Sr;Sv
tel-00785991, version 1 - 7 Feb 2013
normalized reflectance spectrum by dividing the asteroid spectrum
by the solar analog spectrum and performing a correction
for telluric lines.
For the first step, the Image Reduction and Analysis Facility
(Tody 1986) was used. Preprocessing of the CCD images included
bias and flat field correction. An averaged bias frame
taken at the telescope at the beginning of each observing night
was used to perform bias subtraction. Flat fields images were obtained
for each object using calibration lamps, at the beginning
or end of the night. For the wavelength calibration, the Ar lamp
spectrum was used. In the second step, specific IDL routines
were used to help diminish the influence of telluric bands in our
spectra (Rivkin et al. 2004). No other correction for the differential
refraction was performed. For the computation of the final
reflectance (ratio of the asteroid spectrum to the star spectrum),
we took into account the similar dynamic regimes of the detector
(Vacca et al. 2004; Rayner et al. 2003).
Log of asteroids observations is given in Table 1. In general,
the asteroid spectra were obtained taking images with an integration
time (Itime) of 120s in the nodding procedure, for several
cycles, to increase the S/N ratio. For two objects of our sample
(2005 GN59, and 2001 SG286), the atmospheric conditions
and their low brightness imply a poor S/N ratio. In this case, to
obtain reliable spectral measurements, the images were selected
by visual inspection, removing all those in which we could not
distinguish the trace of the spectrum before the data reduction
procedure.
3. Methods used to analyze data
We consider our analysis of spectra in the context of previously
published physical and dynamical properties of these objects.
Table 2 summarizes some parameters of our sample.
We complete our spectral data with the visible counterpart,
when available. This is the case for four of our asteroids: (1917)
Cuyo, (8567) 1996 HW1, (16960) 1998 QS52, and 2001 SG286.
For each of them, the visible spectrum was merged with our
NIR data using a procedure of minimization of data in the common
spectral region 0.82–0.9 μm.
We computed the slope for each spectrum using a first-order
polynomial fit. For the case of composite spectra (V + NIR), the
slope was computed for the spectra normalized to 0.55 μm to
compare with the conclusions of the DeMeo et al. (2009) taxonomy.
Otherwise, when only NIR was available, the slope was
computed for the spectra normalized to 1.25 μm.
Taxonomic types, although not usable to determine the
mineralogic compositions of the objects, help constrain mineral
species that may be present on the surface of the asteroid.
Currently, the most commonly taxonomies are: Tholen
taxonomy (Tholen 1984), Barucci taxonomy (Barucci et al.
1987; Birlan et al. 1996), based on Eight-Color Asteroid Survey
data (Zellner et al. 1985), SMASII spectral taxonomy (Bus &
Binzel 2002), and Bus-DeMeo taxonomy (DeMeo et al. 2009).
We used the last one, which is an extension of the Bus & Binzel
(2002) taxonomy to the near-infrared, considering the data
spanning the wavelength range between 0.45 μm to2.45μm.
The Bus-DeMeo taxonomy is based on Principal Component
Analysis and comprised 24 classes. This taxonomy allows us to
analyze spectra using only NIR data, although we are in this case
unable to obtain a unique classification. We used two independent
methods to establish the taxonomical class of each asteroid
in our sample. In a first approach, spectral data of our asteroids
were compared with Bus-DeMeo taxonomic classes via the
MIT-SMASS on-line tool 2 . The second approach to taxonomic
classification was a procedure developed during this study using
a χ 2 minimization method accounting for the mean and standarddeviation
values of the Bus-DeMeo taxonomic classes. For this
method, we define a reliability criterion:
Reliability = card([λ m,λ M ] ⋂ {λ T 1 ,λT 2 , ..., λT 41 })
(1)
41
where [λ m ,λ M ] is the spectral interval between the minimum and
maximum wavelengths of the spectrum, λ T 1 ,λT 2 , ..., λ 41 are the 41
wavelengths from Bus-DeMeo taxonomy, spanning the interval
between 0.45 μm and 2.45 μm, and card() represents the number
of elements of a discrete set. To apply this procedure, we
smoothed our data by curve fitting with polynomial functions.
This was done using poly fit from the Octave3.2 computation
environment. The degree of the polynomial was selected to be
between 15 to 21 such that the fit produces the smallest least
squares fitting residuals. The obtained fitting curves are given in
Fig. 3. Overall, we observed that both procedures gave similar
results.
We also compared our observational data with laboratory
spectra. Spectroscopy of different samples made in the laboratory
provides the basis upon which compositional information
about unexplored or unsampled planetary surfaces is derived
from remotely obtained reflectance spectra. The Relab 3 spectral
database contains more than 15 000 spectra for different types
of materials from meteorites to terrestrial rocks, man-made mixtures,
and terrestrial and lunar soils. The comparison was made
in the first step using a χ 2 minimization method between asteroid
spectrum and all the spectra from the Relab database, which
were first normalized to 1.25 μm. In this way, the best fifty spectral
curves of different Relab samples were selected for further
analysis. From these results, only the meteorite spectra were
2 http://smass.mit.edu/busdemeoclass.html
3 http://www.planetary.brown.edu/relab/
A15, page 3 of 15